Erasing 'Ugly' from the Internet: Propagation of the Beauty Myth in Text-Image Models

2025-11-01

AI Summary

This study investigates how generative AI systematically encodes Western-centric beauty standards, marginalizing diverse bodily features and exacerbating aesthetic oppression and self-objectification risks for women and non-binary individuals. Methodologically, we design a text-to-image dual-path generation pipeline integrating three language models and two diffusion models, coupled with an in-subject Likert-scale evaluation (N = 1,200 images) co-conducted by women and non-binary participants. Our structured aesthetic taxonomy reveals: (1) negative physical descriptors (e.g., “broad nose”) significantly increase NSFW scores; (2) 86.5% of generated figures exhibit light skin tones, 74% appear youthfully aged, and 22% contain unintended explicit content; (3) non-binary representations are disproportionately youthful and hypersexualized, exposing intersecting age–gender–race biases. Crucially, we provide the first empirical evidence that developers actively suppress bodily diversity via negative prompting—offering both critical evidence and a methodological framework for AI aesthetic governance.

Technology Category

• Natural Language Processing: Generation
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Philosophy and Ethics of AI: Bias, Fairness & Equity

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Social media has exacerbated the promotion of Western beauty norms, leading to negative self-image, particularly in women and girls, and causing harm such as body dysmorphia. Increasingly content on the internet has been artificially generated, leading to concerns that these norms are being exaggerated. The aim of this work is to study how generative AI models may encode 'beauty' and erase 'ugliness', and discuss the implications of this for society. To investigate these aims, we create two image generation pipelines: a text-to-image model and a text-to-language model-to image model. We develop a structured beauty taxonomy which we use to prompt three language models (LMs) and two text-to-image models to cumulatively generate 5984 images using our two pipelines. We then recruit women and non-binary social media users to evaluate 1200 of the images through a Likert-scale within-subjects study. Participants show high agreement in their ratings. Our results show that 86.5% of generated images depicted people with lighter skin tones, 22% contained explicit content despite Safe for Work (SFW) training, and 74% were rated as being in a younger age demographic. In particular, the images of non-binary individuals were rated as both younger and more hypersexualised, indicating troubling intersectional effects. Notably, prompts encoded with 'negative' or 'ugly' beauty traits (such as "a wide nose") consistently produced higher Not SFW (NSFW) ratings regardless of gender. This work sheds light on the pervasive demographic biases related to beauty standards present in generative AI models -- biases that are actively perpetuated by model developers, such as via negative prompting. We conclude by discussing the implications of this on society, which include pollution of the data streams and active erasure of features that do not fall inside the stereotype of what is considered beautiful by developers.

Problem

• Examining how AI models encode beauty standards and erase ugliness
• Investigating demographic biases in generative AI image outputs
• Analyzing societal implications of perpetuated beauty stereotypes

Innovation

• Developed structured beauty taxonomy for model prompts
• Created dual text-to-image generation pipelines
• Used human evaluation to quantify demographic biases